This paper presents applications of multiscale mathematical morphology to contrast enhancement of medical, infrared, visible images and its extension to visible and infrared images fusion. Classic top-hat transform is a mathematical morphological operation that is used in contrast enhancement and images fusion. But this operation introduces saturations into the image enhancement process. Therefore, a generalization of the top-hat transform using two structuring elements is proposed. This variation allows to decrease the saturations in the images processed. Also, this variation is tested in new algorithms based on multiscale mathematical morphology applied to contrast enhancement of medical, infrared and visible images and in a new algorithm based on multiscale mathematical morphology for visible and infrared images fusion. The results obtained show medical, infrared, visible and fused images with contrast enhancement and detail improvement.